Nouveau - Oct 2014 - v3.17, i915 vs nouveau: possible recursive locking detected

============================================[ INFO: possible recursive locking
detected ]
3.17.0 #50 Not tainted
---------------------------------------------
Xorg/1170 is trying to acquire lock:
 (&dev->struct_mutex){+.+.+.}, at: [<ffffffffa010ae93>]
i915_gem_unmap_dma_buf+0x33/0xc0 [i915]

but task is already holding lock:
 (&dev->struct_mutex){+.+.+.}, at: [<ffffffffa001cd1a>]
drm_gem_object_handle_unreference_unlocked+0xca/0x120 [drm]

other info that might help us debug this:
 Possible unsafe locking scenario:

       CPU0
       ----
  lock(&dev->struct_mutex);
  lock(&dev->struct_mutex);

 *** DEADLOCK ***

 May be due to missing lock nesting notation

1 lock held by Xorg/1170:
 #0:  (&dev->struct_mutex){+.+.+.}, at: [<ffffffffa001cd1a>]
drm_gem_object_handle_unreference_unlocked+0xca/0x120 [drm]

stack backtrace:
CPU: 1 PID: 1170 Comm: Xorg Not tainted 3.17.0 #50
Hardware name: Dell Inc.          Inspiron 7720/04M3YM, BIOS A07 08/16/2012
 ffffffff82606b80 ffff880136e27b00 ffffffff81707891 ffffffff82606b80
 ffff880136e27bb8 ffffffff810e0e4f ffffffff00000000 000202000000c795
 0000000000000001 0000000000000000 00000000004ec276 0000000000000300
Call Trace:
 [<ffffffff81707891>] dump_stack+0x4e/0x7a
 [<ffffffff810e0e4f>] __lock_acquire+0x194f/0x19e0
 [<ffffffff810e1631>] lock_acquire+0x91/0x120
 [<ffffffffa010ae93>] ? i915_gem_unmap_dma_buf+0x33/0xc0 [i915]
 [<ffffffffa010ae93>] ? i915_gem_unmap_dma_buf+0x33/0xc0 [i915]
 [<ffffffff8170c014>] mutex_lock_nested+0x54/0x3d0
 [<ffffffffa010ae93>] ? i915_gem_unmap_dma_buf+0x33/0xc0 [i915]
 [<ffffffff810df03a>] ? mark_held_locks+0x6a/0x90
 [<ffffffffa010ae93>] i915_gem_unmap_dma_buf+0x33/0xc0 [i915]
 [<ffffffff814c3032>] dma_buf_unmap_attachment+0x22/0x40
 [<ffffffffa0034e42>] drm_prime_gem_destroy+0x22/0x40 [drm]
 [<ffffffffa0299b5b>] nouveau_gem_object_del+0x3b/0x60 [nouveau]
 [<ffffffffa001c7b7>] drm_gem_object_free+0x27/0x30 [drm]
 [<ffffffffa001cd34>]
drm_gem_object_handle_unreference_unlocked+0xe4/0x120 [drm]
 [<ffffffffa001ce2a>] drm_gem_handle_delete+0xba/0x110 [drm]
 [<ffffffffa001d495>] drm_gem_close_ioctl+0x25/0x30 [drm]
 [<ffffffffa001df0c>] drm_ioctl+0x1ec/0x660 [drm]
 [<ffffffff8148e4b2>] ? __pm_runtime_resume+0x32/0x60
 [<ffffffff817102fd>] ? _raw_spin_unlock_irqrestore+0x5d/0x70
 [<ffffffff810df15d>] ? trace_hardirqs_on_caller+0xfd/0x1c0
 [<ffffffff810df22d>] ? trace_hardirqs_on+0xd/0x10
 [<ffffffff817102e2>] ? _raw_spin_unlock_irqrestore+0x42/0x70
 [<ffffffffa0290bd4>] nouveau_drm_ioctl+0x54/0xc0 [nouveau]
 [<ffffffff812072a0>] do_vfs_ioctl+0x300/0x520
 [<ffffffff81211fad>] ? __fget+0xad/0xe0
 [<ffffffff81211f05>] ? __fget+0x5/0xe0
 [<ffffffff81207541>] SyS_ioctl+0x81/0xa0
 [<ffffffff81710ed6>] system_call_fastpath+0x1a/0x1f

We need ww mutexes and need to rewrite i915 a bit fo fix this all. I.e.
known issue. As long as your userspace isn't nasty nothing bad will ever
happen though.
-Daniel

On Mon, Oct 13, 2014 at 08:40:33PM +0200, Marcin ?lusarz
wrote:
> ============================================> [ INFO: possible recursive
locking detected ]
> 3.17.0 #50 Not tainted
> ---------------------------------------------
> Xorg/1170 is trying to acquire lock:
>  (&dev->struct_mutex){+.+.+.}, at: [<ffffffffa010ae93>]
i915_gem_unmap_dma_buf+0x33/0xc0 [i915]
> 
> but task is already holding lock:
>  (&dev->struct_mutex){+.+.+.}, at: [<ffffffffa001cd1a>]
drm_gem_object_handle_unreference_unlocked+0xca/0x120 [drm]
> 
> other info that might help us debug this:
>  Possible unsafe locking scenario:
> 
>        CPU0
>        ----
>   lock(&dev->struct_mutex);
>   lock(&dev->struct_mutex);
> 
>  *** DEADLOCK ***
> 
>  May be due to missing lock nesting notation
> 
> 1 lock held by Xorg/1170:
>  #0:  (&dev->struct_mutex){+.+.+.}, at: [<ffffffffa001cd1a>]
drm_gem_object_handle_unreference_unlocked+0xca/0x120 [drm]
> 
> stack backtrace:
> CPU: 1 PID: 1170 Comm: Xorg Not tainted 3.17.0 #50
> Hardware name: Dell Inc.          Inspiron 7720/04M3YM, BIOS A07 08/16/2012
>  ffffffff82606b80 ffff880136e27b00 ffffffff81707891 ffffffff82606b80
>  ffff880136e27bb8 ffffffff810e0e4f ffffffff00000000 000202000000c795
>  0000000000000001 0000000000000000 00000000004ec276 0000000000000300
> Call Trace:
>  [<ffffffff81707891>] dump_stack+0x4e/0x7a
>  [<ffffffff810e0e4f>] __lock_acquire+0x194f/0x19e0
>  [<ffffffff810e1631>] lock_acquire+0x91/0x120
>  [<ffffffffa010ae93>] ? i915_gem_unmap_dma_buf+0x33/0xc0 [i915]
>  [<ffffffffa010ae93>] ? i915_gem_unmap_dma_buf+0x33/0xc0 [i915]
>  [<ffffffff8170c014>] mutex_lock_nested+0x54/0x3d0
>  [<ffffffffa010ae93>] ? i915_gem_unmap_dma_buf+0x33/0xc0 [i915]
>  [<ffffffff810df03a>] ? mark_held_locks+0x6a/0x90
>  [<ffffffffa010ae93>] i915_gem_unmap_dma_buf+0x33/0xc0 [i915]
>  [<ffffffff814c3032>] dma_buf_unmap_attachment+0x22/0x40
>  [<ffffffffa0034e42>] drm_prime_gem_destroy+0x22/0x40 [drm]
>  [<ffffffffa0299b5b>] nouveau_gem_object_del+0x3b/0x60 [nouveau]
>  [<ffffffffa001c7b7>] drm_gem_object_free+0x27/0x30 [drm]
>  [<ffffffffa001cd34>]
drm_gem_object_handle_unreference_unlocked+0xe4/0x120 [drm]
>  [<ffffffffa001ce2a>] drm_gem_handle_delete+0xba/0x110 [drm]
>  [<ffffffffa001d495>] drm_gem_close_ioctl+0x25/0x30 [drm]
>  [<ffffffffa001df0c>] drm_ioctl+0x1ec/0x660 [drm]
>  [<ffffffff8148e4b2>] ? __pm_runtime_resume+0x32/0x60
>  [<ffffffff817102fd>] ? _raw_spin_unlock_irqrestore+0x5d/0x70
>  [<ffffffff810df15d>] ? trace_hardirqs_on_caller+0xfd/0x1c0
>  [<ffffffff810df22d>] ? trace_hardirqs_on+0xd/0x10
>  [<ffffffff817102e2>] ? _raw_spin_unlock_irqrestore+0x42/0x70
>  [<ffffffffa0290bd4>] nouveau_drm_ioctl+0x54/0xc0 [nouveau]
>  [<ffffffff812072a0>] do_vfs_ioctl+0x300/0x520
>  [<ffffffff81211fad>] ? __fget+0xad/0xe0
>  [<ffffffff81211f05>] ? __fget+0x5/0xe0
>  [<ffffffff81207541>] SyS_ioctl+0x81/0xa0
>  [<ffffffff81710ed6>] system_call_fastpath+0x1a/0x1f
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch

On Thu, 16 Oct 2014 13:47:38 +0200
Daniel Vetter <daniel at ffwll.ch> wrote:
> We need ww mutexes and need to rewrite i915 a bit fo fix this all.
> I.e. known issue. As long as your userspace isn't nasty nothing bad
> will ever happen though.
So do we already have a bug open with a good description of the issue?
Eventually the streets will run with blood and we'll enter the 6th
republic of i915, so it would be good to have it on the hit list for
when the time comes...

Thanks,
Jesse